كاربرد منطق فازي جهت برآورد تبخير - تعرق پتانسيل شهرستان گناباد

Use of Fuzzy Logic to Determine the Evapotranspiration in Gonabad City

زﻣﯿﻨﻪ و ﻫﺪف: ﺗﺒﺨﯿﺮ- ﺗﻌﺮق ﭘﺘﺎﻧﺴﯿﻞ ﯾﮑﯽ از اﺟﺰاي اﺻﻠﯽ ﭼﺮﺧﻪ ي ﻫﯿﺪروﻟﻮژﯾﮑﯽ اﺳﺖ ﮐﻪ ﺗﻌﯿﯿﻦ ﺻﺤﯿﺢ آن در ﻣﻄﺎﻟﻌﺎت ﺑﯿﻼن آﺑﯽ، ﻃﺮاﺣﯽ ﺳﯿﺴﺘﻢ ﻫﺎي آﺑﯿﺎري و ﺑﺮﻧﺎﻣﻪ رﯾﺰي و ﻣﺪﯾﺮﯾﺖ ﻣﻨﺎﺑﻊ آب ﺑﺮاي دﺳﺖ ﯾﺎﺑﯽ ﺑﻪ ﺗﻮﺳﻌﻪ ي ﭘﺎﯾﺪار ﻧﻘﺶ ﺑﻪ ﺳﺰاﯾﯽ دارد. ﺗﺒﺨﯿﺮ و ﺗﻌﺮق ﺑﻪ ﻋﻠﺖ ﻧﯿﺎز ﺑﻪ ﻋﻮاﻣﻞ اﻗﻠﯿﻤﯽ ﻣﺨﺘﻠﻒ و اﺛﺮ ﻣﺘﻘﺎﺑﻞ اﯾﻦ ﻋﻮاﻣﻞ ﺑﺮ ﯾﮑﺪﯾﮕﺮ ﯾﮏ ﭘﺪﯾﺪه ي ﻏﯿﺮﺧﻄﯽ و ﭘﯿﭽﯿﺪه اﺳﺖ. ﯾﮑﯽ از ﻣﺮاﺣﻞ ﭘﯿﭽﯿﺪه در ﻣﺪل ﺳﺎزي ﺳﯿﺴﺘﻢ ﻫﺎي ﻏﯿﺮﺧﻄﯽ، ﭘﯿﺶ ﭘﺮدازش ﭘﺎراﻣﺘﺮ ﻫﺎي ورودي ﺟﻬﺖ اﻧﺘﺨﺎب ﺗﺮﮐﯿﺒﯽ ﻣﻨﺎﺳﺐ از آن ﻫﺎﺳﺖ. ﭘﯿﺶ ﭘﺮدازش داده ﻫﺎ ﺳﺒﺐ ﮐﺎﻫﺶ ﻣﺮاﺣﻞ ﺳﻌﯽ و ﺧﻄﺎ و ﺷﻨﺎﺧﺖ ﻣﻬﻢﺗﺮﯾﻦ ﭘﺎراﻣﺘﺮﻫﺎي ﻣﺆﺛﺮ ﺑﺮ ﭘﺪﯾﺪه ي ﻣﻮرد ﻧﻈﺮ ﺟﻬﺖ ﻣﺪل ﺳﺎزي ﺑﺎ اﺳﺘﻔﺎده از روش ﻫﺎي ﻫﻮش ﻣﻨﺪ ﻣﯽﮔﺮدد. اﯾﻦ ﭘﮋوﻫﺶ ﺑﺎ ﻫﺪف اﺳﺘﻔﺎده از ﺗﻮان ﻣﻨﺪي ﻫﺎي ﺳﯿﺴﺘﻢ اﺳﺘﻨﺘﺎج ﻓﺎزي ﺑﺮاي ﺑﺮآورد ﺗﺒﺨﯿﺮ و ﺗﻌﺮق ﺑﺎ اﺳﺘﻔﺎده از داده ﻫﺎي ﻫﻮاﺷﻨﺎﺳﯽ ﻃﯽ دوره 20 ﺳﺎﻟﻪ )1372-1392( در ﺷﻬﺮ ﮔﻨﺎﺑﺎد ﺻﻮرت ﮔﺮﻓﺖ. روش ﺑﺮرﺳﯽ: ﺑﺮاﯾﻦ اﺳﺎس ﭘﺲ از ﺑﺮرﺳﯽ ﻣﺪل ﻣﻮﺟﻮد و ﺑﺮرﺳﯽ ﺗﺮﮐﯿﺐﻫﺎي ﻣﺨﺘﻠﻒ داده ﻫﺎي ﻫﻮاﺷﻨﺎﺳﯽ، ﻣﺪل ﻧﻬﺎﯾﯽ ﺑﺮاي ﺑﺮآورد ﺗﺒﺨﯿﺮ و ﺗﻌﺮق اراﯾﻪ ﺷﺪ. در اﯾﻦ ﺗﺤﻘﯿﻖ در ﻣﺠﻤﻮع ﺑﺎ داﺷﺘﻦ 20 ﻣﺘﻐﯿﺮ ﺑﺮاي ورودي ﻣﺪل و ﯾﮏ ﻣﺘﻐﯿﺮ ﺑﺮاي ﺧﺮوﺟﯽ ﻣﺪل )ﺗﺒﺨﯿﺮ و ﺗﻌﺮق(، 50 ﻗﺎﻧﻮن در ﺳﯿﺴﺘﻢ اﺳﺘﻨﺘﺎج ﻣﻤﺪاﻧﯽ ﺗﻌﺮﯾﻒ ﺷﺪ. ﻣﻘﺎدﯾﺮ ﺗﺒﺨﯿﺮ و ﺗﻌﺮق ﭘﺘﺎﻧﺴﯿﻞ ﺣﺎﺻﻞ از ﻣﺪل ﻓﺎزي ﺑﺎ ﻣﺪل ﭘﻨﻤﻦ ﻓﺎﺋﻮ- ﻣﻮﻧﺘﯿﺚ ﻣﻮرد ﻣﻘﺎﯾﺴﻪ ﻗﺮار ﮔﺮﻓﺖ. ﺑﺤﺚ و ﻧﺘﯿﺠﻪ ﮔﯿﺮي: ﮐﺎراﯾﯽ ﻣﺪل ﻓﺎزي اراﯾﻪ ﺷﺪه ﺑﺎ اﺳﺘﻔﺎده از آﻣﺎره ﻫﺎي رﯾﺸﻪ ﻣﯿﺎﻧﮕﯿﻦ ﻣﺮﺑﻌﺎت ﺧﻄﺎ، ﺧﻄﺎي اﻧﺤﺮاف ﻣﯿﺎﻧﮕﯿﻦ، ﺿﺮﯾﺐ ﺗﻌﯿﯿﻦ و ﻣﻌﯿﺎر ﺟﺎﮐﻮوﯾﺪز )t( و ﻣﻌﯿﺎر ﺻﺒﺎغ و ﻫﻤﮑﺎران )(R2 /t ﻣﻮرد ارزﯾﺎﺑﯽ ﻗﺮار ﮔﺮﻓﺖ. ﻣﻘﺎﯾﺴﻪ ﻧﺘﺎﯾﺞ ﻣﺪل ﻓﺎزي ﺑﺎ ﻣﺪل ﭘﻨﻤﻦ ﻓﺎﺋﻮ- ﻣﻮﻧﺘﯿﺚ ﻣﻘﺎﯾﺴﻪ ﮔﺮدﯾﺪ. ﺑﯿﺶ ﺗﺮﯾﻦ ﺗﺒﺨﯿﺮ و ﺗﻌﺮق در ﻣﺎه ژوﺋﯿﻪ اﺗﻔﺎق اﻓﺘﺎده اﺳﺖ. ﻧﺘﺎﯾﺞ ﺣﺎﮐﯽ از ﻫﻤﺒﺴﺘﮕﯽ ﺑﺎﻻ ﺑﯿﻦ اﯾﻦ دو ﻣﺪل ﻓﺎزي اراﯾﻪ ﺷﺪه و روش ﭘﻨﻤﻦ ﻓﺎﺋﻮ- ﻣﻮﻧﺘﯿﺚ اﺳﺖ.

Background and Objectives: Evapotranspiration are the main component of hydrologic cycle and estimation of the amount of evapotranspiration is important to study the water balance, design the irrigation systems and ultimately plan and manage the water resources to achieve the stable development. Evapotranspiration are nonlinear and complicated phenomenon due to requirement for different factors and their interactions. One important step in non-linear system modeling is pre- investigation of inputs to achieve suitable combination of them. Pre-investigation of input data prevents several trial and error steps and helps to understand the most important parameters which affect the phenomenon to be able to modeling the system in an intelligent way. Thereby, this study, has aimed to implement the ability of Fuzzy logic system to estimate Evapotranspiration by using the data from Gonabad weather station in a 21-year period (1993-2014). Material and Methods: To reach this aim, after investigation of available models and different combination of weather information, the final model to estimate Evapotranspiration has been designed. In this model with 20 surface as input and one surface as output or evaporation and transpiration, 50 rules were determined in Mamdani Inference System and the estimated value of Evapotranspiration from the Fuzzy Inference were compared with the results from Fao Penman Monteith (F-P-M). Results and Discussion: Statistical parameters including Root Mean Square Error (RMSE), Main Bias Error (MBE) Coefficient of Determination, Jacovides factor (t) and Sabbagh et al. factor (R2/t) have been used to investigate the efficiency of this model. Comparison of the results from the Fuzzy Model and the results from F-P-M shows a high level of correlation between these two methods indicating the highest level of Evapotranspiration in July (RMSE: 0.08, R2: 0.98, MBE: 0.05, t:0.25 and R2/t: 3.92).